Image Compression and Decompression Framework Based on Latent Diffusion Model for Breast Mammography

TL;DR

This paper introduces a novel medical image compression framework using Latent Diffusion Models, achieving superior quality, reduced dataset size, and comparable diagnostic performance, with potential applications in noise reduction and efficient storage.

Contribution

The study develops a new compression-decompression framework based on Latent Diffusion Models that outperforms traditional algorithms in medical imaging.

Findings

Surpasses conventional file compression algorithms in image quality.

CNN models trained on decompressed images perform comparably to those trained on original images.

Significantly reduces dataset size for easier distribution and storage.

Abstract

This research presents a novel framework for the compression and decompression of medical images utilizing the Latent Diffusion Model (LDM). The LDM represents advancement over the denoising diffusion probabilistic model (DDPM) with a potential to yield superior image quality while requiring fewer computational resources in the image decompression process. A possible application of LDM and Torchvision for image upscaling has been explored using medical image data, serving as an alternative to traditional image compression and decompression algorithms. The experimental outcomes demonstrate that this approach surpasses a conventional file compression algorithm, and convolutional neural network (CNN) models trained with decompressed files perform comparably to those trained with original image files. This approach also significantly reduces dataset size so that it can be distributed with a smaller size, and medical images take up much less space in medical devices. The research implications extend to noise reduction in lossy compression algorithms and substitute for complex wavelet-based lossless algorithms.